Abstract Details

Neurologic infections remain a global contributor to morbidity and mortality. In many low-income countries, despite its low cost, manual microscopy, the gold standard for measuring CSF WBC remains limited or unreliable. CSF-specific modes for hematologic analyzers have emerged as a less skill-dependent alternative.

This systematic review was conducted following the Cochrane Collaboration guidelines. Assessment of Bias was performed using QUADAS-2 Tool. Data were pooled and analyzed by method of comparison using fischer-z transformation for correlation coefficients and inverse logit transformations for proportions. All pooled estimated used random effects estimation.

Among 652 titles, 554 abstract screened, 104 underwent full text review, and 111 comparisons from 41 studies were included. Among these studies, 37 (90%) were performed in high-income countries. There were 51 automated analyzers included reflecting nine brands, with Sysmex brand the most common 27 (53%). Pooled estimates of sensitivity and specificity (n=7) were 95 (95%-CI  93–97%) and 84% (95%-CI: 64-96%), respectively. Pooled R² estimates (n=29) were 0.95 (95%-CI: 0.95-0.96); Pooled spearman rho correlation (n=27) estimates were 0.95 (95% CI 0.95-0.96). Among comparisons using Passing Bablok regressions (n=14) the pooled slope was estimated to be 1.05 (95% CI 1.03-1.07). Among those comparisons using Bland-Altman plot (n=11) pooled mean difference was estimated at 0.98 (95% CI -0.54–2.5). Q tests of homogeneity were all significant with the exception of the Bland-altman comparisons (I² 10%, p value 0.35).